9
$\begingroup$

The tag [mapreduce] doesn't have a user guidance. Currently it has questions that cover both the technique of "Map-Reduce" from concurrent algorithms and "map reduction" from theory of computability.

Currently it has only 11 questions, most of them about concurrent programming. So I suggest:

  1. remove the tag; or
  2. write a user guide specifying the tag is about concurrent programming, and remove the tag from questions about computer theory.

Either fixes would be very easy. However I would prefer option #2. This tag has been seldom used, but this might be because it lacks user guidance.

$\endgroup$
  • $\begingroup$ What does "about concurrent programming" mean? We're not a programming site, so it sounds like your proposal is to remove the tag from the on-topic questions that use it, and reserve the tag only for off-topic questions. $\endgroup$ – David Richerby Mar 1 '18 at 9:32
  • $\begingroup$ I suppose a few of the questions under that tag are about the concept of map-reduce, not Hadoop, CouchDB or any specific implementation. If this still would be off-topic, then I guess the tag should be removed. There is already a tag for language reduction and if we were to have tags about specific types of reduction, they should be [cook-reduction] and [many-one-reduction]/[karp-reduction], not [map-reduce], to avoid confusion. $\endgroup$ – giusti Mar 1 '18 at 12:35
  • 1
    $\begingroup$ The concept is fine, but anything relating to a specific implementation is off-topic. If people are using the map-reduce tag for questions about mapping reductions, that's just wrong. If the only actual problem is that people are doing that, then those questions should be retagged. $\endgroup$ – David Richerby Mar 1 '18 at 12:40
  • $\begingroup$ I think the tag does not necessarily need be specific to an implementation. Map-reduce started as a Google implementation, but is now a broader concept for algorithms that, well, map, and then reduce! So the tag might have merit. But really, I'm fine with any solution that the community decide. $\endgroup$ – giusti Mar 1 '18 at 12:55
  • 1
    $\begingroup$ More strongly, the tag needs not to be specific to any implementation. $\endgroup$ – David Richerby Mar 1 '18 at 13:00
5
$\begingroup$

Thanks for bringing this up!

Since you already have a good overview, please

If you can't access that link due to missing privileges, please provide a candidate text here.

$\endgroup$
  • $\begingroup$ OK, I started with a basic draft. This is sufficient to distinguish from computability and separate the questions. But I can't define the whole scope alone. For example, I can only guess that questions specific to implementation should go to StackOverflow, and so on. $\endgroup$ – giusti Feb 28 '18 at 16:35
  • $\begingroup$ I voted to reject both tag edits, since tag descriptions are supposed to explain what the tag is used for, not just define the term in the tag's name. $\endgroup$ – David Richerby Mar 1 '18 at 9:31
  • 1
    $\begingroup$ That's not very helpful. I won't write the whole tag description by myself! Starting with a draft at least makes the focus of the tag clear and allows us to sort the questions accordingly. $\endgroup$ – giusti Mar 1 '18 at 12:38
  • 1
    $\begingroup$ @giusti Then please post your draft as an answer here. Edit proposals are ill suited for discussion and collaboration. $\endgroup$ – Raphael Mar 1 '18 at 12:40
  • 1
    $\begingroup$ @DavidRicherby I don't know if that's available for tag edits, but "improve edit" might have been the more positive action to take. $\endgroup$ – Raphael Mar 1 '18 at 12:41
  • $\begingroup$ You're right, my tag suggestion does not cover all the points on the sidebar. I just wanted to make sure the scope is defined before writing anything that might be removed later. $\endgroup$ – giusti Mar 1 '18 at 12:45
  • $\begingroup$ Can I withdraw my request for tag creation? I just wanted to copy the text and I think I accidentally submitted a useless request for tag creation... Sorry. $\endgroup$ – giusti Mar 1 '18 at 12:46
  • $\begingroup$ @Raphael I'd have improved the edit if I had improvements, but recognizing that something is wrong requires less skill than fixing it, and I don't know anything like enough about map-reduce to write tag entries for it. $\endgroup$ – David Richerby Mar 1 '18 at 12:56
  • $\begingroup$ @giusti I rejected the proposal, so we can polish it here first. No biggy. $\endgroup$ – Raphael Mar 1 '18 at 12:57
  • $\begingroup$ @DavidRicherby Fair enough. $\endgroup$ – Raphael Mar 1 '18 at 12:58
1
$\begingroup$

This is the initial proposition for the tag excerpt.

Map reduce is a programming model for parallel programming. Questions about parallel architecture and speed-up analysis using map-reduce should be tagged here. Questions about programming, or about specific implementations of map-reduce (Hadoop, CouchDB, etc) are off-topic here.

Once the scope of the tag has been approved, the full tag description should contain:

► what questions should have this tag?

► some basic definitions

► brief introduction to the subject

► important links for learning more

► one reasonably sized page

This is not an answer to the question, but a draft posted as an answer by request of a moderator.

$\endgroup$
  • 2
    $\begingroup$ Once "[...]" is fleshed out, I think that will be a good tag excerpt. $\endgroup$ – David Richerby Mar 1 '18 at 12:57
  • $\begingroup$ Is it about concurrent or parallel programming? I think the latter. $\endgroup$ – Raphael Mar 1 '18 at 12:58
  • $\begingroup$ @Raphael Told you I didn't know enough about map-reduce! :-D $\endgroup$ – David Richerby Mar 1 '18 at 12:58
  • $\begingroup$ I don't know how to finish this. I can only give one example of a kind of question that would be eligible: when trying to calculate the speedup of an algorithm modelled with map-reduce. But I don't believe that's all. I don't understand the subject well enough to properly define its scope. $\endgroup$ – giusti Mar 6 '18 at 13:14
  • $\begingroup$ OK, I took a "best-effort" approach using the general guidelines of the CS StackExchange and wrote something. Let me know if it's adequate. $\endgroup$ – giusti Mar 6 '18 at 13:21

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .